DEPR: Series/DataFrame with tzaware data and tznaive dtype (pandas-dev#41555)

Author: jbrockmendel

doc/source/whatsnew/v1.3.0.rst

@@ -677,6 +677,7 @@ Deprecations
 - Deprecated special treatment of lists with first element a Categorical in the :class:`DataFrame` constructor; pass as ``pd.DataFrame({col: categorical, ...})`` instead (:issue:`38845`)
 - Deprecated passing arguments as positional (except for ``"method"``) in :meth:`DataFrame.interpolate` and :meth:`Series.interpolate` (:issue:`41485`)
 - Deprecated passing arguments (apart from ``value``) as positional in :meth:`DataFrame.fillna` and :meth:`Series.fillna` (:issue:`41485`)
+- Deprecated construction of :class:`Series` or :class:`DataFrame` with ``DatetimeTZDtype`` data and ``datetime64[ns]`` dtype.  Use ``Series(data).dt.tz_localize(None)`` instead (:issue:`41555`,:issue:`33401`)
 
 .. ---------------------------------------------------------------------------
 

pandas/core/dtypes/cast.py

@@ -220,6 +220,8 @@ def maybe_unbox_datetimelike(value: Scalar, dtype: DtypeObj) -> Scalar:
     elif isinstance(value, Timestamp):
         if value.tz is None:
             value = value.to_datetime64()
+        elif not isinstance(dtype, DatetimeTZDtype):
+            raise TypeError("Cannot unbox tzaware Timestamp to tznaive dtype")
     elif isinstance(value, Timedelta):
         value = value.to_timedelta64()
 
@@ -1628,9 +1630,21 @@ def maybe_cast_to_datetime(
                             # didn't specify one
 
                             if dta.tz is not None:
+                                warnings.warn(
+                                    "Data is timezone-aware. Converting "
+                                    "timezone-aware data to timezone-naive by "
+                                    "passing dtype='datetime64[ns]' to "
+                                    "DataFrame or Series is deprecated and will "
+                                    "raise in a future version. Use "
+                                    "`pd.Series(values).dt.tz_localize(None)` "
+                                    "instead.",
+                                    FutureWarning,
+                                    stacklevel=8,
+                                )
                                 # equiv: dta.view(dtype)
                                 # Note: NOT equivalent to dta.astype(dtype)
                                 dta = dta.tz_localize(None)
+
                             value = dta
                         elif is_datetime64tz:
                             dtype = cast(DatetimeTZDtype, dtype)
@@ -1822,7 +1836,7 @@ def construct_2d_arraylike_from_scalar(
     shape = (length, width)
 
     if dtype.kind in ["m", "M"]:
-        value = maybe_unbox_datetimelike(value, dtype)
+        value = maybe_unbox_datetimelike_tz_deprecation(value, dtype, stacklevel=4)
     elif dtype == object:
         if isinstance(value, (np.timedelta64, np.datetime64)):
             # calling np.array below would cast to pytimedelta/pydatetime
@@ -1885,14 +1899,48 @@ def construct_1d_arraylike_from_scalar(
             if not isna(value):
                 value = ensure_str(value)
         elif dtype.kind in ["M", "m"]:
-            value = maybe_unbox_datetimelike(value, dtype)
+            value = maybe_unbox_datetimelike_tz_deprecation(value, dtype)
 
         subarr = np.empty(length, dtype=dtype)
         subarr.fill(value)
 
     return subarr
 
 
+def maybe_unbox_datetimelike_tz_deprecation(
+    value: Scalar, dtype: DtypeObj, stacklevel: int = 5
+):
+    """
+    Wrap maybe_unbox_datetimelike with a check for a timezone-aware Timestamp
+    along with a timezone-naive datetime64 dtype, which is deprecated.
+    """
+    # Caller is responsible for checking dtype.kind in ["m", "M"]
+    try:
+        value = maybe_unbox_datetimelike(value, dtype)
+    except TypeError:
+        if (
+            isinstance(value, Timestamp)
+            and value.tz is not None
+            and isinstance(dtype, np.dtype)
+        ):
+            warnings.warn(
+                "Data is timezone-aware. Converting "
+                "timezone-aware data to timezone-naive by "
+                "passing dtype='datetime64[ns]' to "
+                "DataFrame or Series is deprecated and will "
+                "raise in a future version. Use "
+                "`pd.Series(values).dt.tz_localize(None)` "
+                "instead.",
+                FutureWarning,
+                stacklevel=stacklevel,
+            )
+            new_value = value.tz_localize(None)
+            return maybe_unbox_datetimelike(new_value, dtype)
+        else:
+            raise
+    return value
+
+
 def construct_1d_object_array_from_listlike(values: Sized) -> np.ndarray:
     """
     Transform any list-like object in a 1-dimensional numpy array of object

pandas/tests/frame/test_constructors.py

@@ -2432,6 +2432,17 @@ def test_from_series_with_name_with_columns(self):
         expected = DataFrame(columns=["bar"])
         tm.assert_frame_equal(result, expected)
 
+    def test_nested_list_columns(self):
+        # GH 14467
+        result = DataFrame(
+            [[1, 2, 3], [4, 5, 6]], columns=[["A", "A", "A"], ["a", "b", "c"]]
+        )
+        expected = DataFrame(
+            [[1, 2, 3], [4, 5, 6]],
+            columns=MultiIndex.from_tuples([("A", "a"), ("A", "b"), ("A", "c")]),
+        )
+        tm.assert_frame_equal(result, expected)
+
 
 class TestDataFrameConstructorWithDatetimeTZ:
     @pytest.mark.parametrize("tz", ["US/Eastern", "dateutil/US/Eastern"])
@@ -2464,12 +2475,41 @@ def test_construction_preserves_tzaware_dtypes(self, tz):
         tm.assert_series_equal(result, expected)
 
     def test_constructor_data_aware_dtype_naive(self, tz_aware_fixture):
-        # GH#25843
+        # GH#25843, GH#41555, GH#33401
         tz = tz_aware_fixture
-        result = DataFrame({"d": [Timestamp("2019", tz=tz)]}, dtype="datetime64[ns]")
-        expected = DataFrame({"d": [Timestamp("2019")]})
+        ts = Timestamp("2019", tz=tz)
+        ts_naive = Timestamp("2019")
+
+        with tm.assert_produces_warning(FutureWarning):
+            result = DataFrame({0: [ts]}, dtype="datetime64[ns]")
+
+        expected = DataFrame({0: [ts_naive]})
+        tm.assert_frame_equal(result, expected)
+
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            result = DataFrame({0: ts}, index=[0], dtype="datetime64[ns]")
+        tm.assert_frame_equal(result, expected)
+
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            result = DataFrame([ts], dtype="datetime64[ns]")
+        tm.assert_frame_equal(result, expected)
+
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            result = DataFrame(np.array([ts], dtype=object), dtype="datetime64[ns]")
+        tm.assert_frame_equal(result, expected)
+
+        with tm.assert_produces_warning(FutureWarning):
+            result = DataFrame(ts, index=[0], columns=[0], dtype="datetime64[ns]")
+        tm.assert_frame_equal(result, expected)
+
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            df = DataFrame([Series([ts])], dtype="datetime64[ns]")
         tm.assert_frame_equal(result, expected)
 
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            df = DataFrame([[ts]], columns=[0], dtype="datetime64[ns]")
+        tm.assert_equal(df, expected)
+
     def test_from_dict(self):
 
         # 8260
@@ -2710,13 +2750,15 @@ def test_from_out_of_bounds_timedelta(self, constructor, cls):
 
         assert type(get1(result)) is cls
 
-    def test_nested_list_columns(self):
-        # GH 14467
-        result = DataFrame(
-            [[1, 2, 3], [4, 5, 6]], columns=[["A", "A", "A"], ["a", "b", "c"]]
-        )
-        expected = DataFrame(
-            [[1, 2, 3], [4, 5, 6]],
-            columns=MultiIndex.from_tuples([("A", "a"), ("A", "b"), ("A", "c")]),
-        )
-        tm.assert_frame_equal(result, expected)
+    def test_tzaware_data_tznaive_dtype(self, constructor):
+        tz = "US/Eastern"
+        ts = Timestamp("2019", tz=tz)
+        ts_naive = Timestamp("2019")
+
+        with tm.assert_produces_warning(
+            FutureWarning, match="Data is timezone-aware", check_stacklevel=False
+        ):
+            result = constructor(ts, dtype="M8[ns]")
+
+        assert np.all(result.dtypes == "M8[ns]")
+        assert np.all(result == ts_naive)

pandas/tests/series/test_constructors.py

@@ -1536,10 +1536,26 @@ def test_constructor_tz_mixed_data(self):
         tm.assert_series_equal(result, expected)
 
     def test_constructor_data_aware_dtype_naive(self, tz_aware_fixture):
-        # GH#25843
+        # GH#25843, GH#41555, GH#33401
         tz = tz_aware_fixture
-        result = Series([Timestamp("2019", tz=tz)], dtype="datetime64[ns]")
-        expected = Series([Timestamp("2019")])
+        ts = Timestamp("2019", tz=tz)
+        ts_naive = Timestamp("2019")
+
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            result = Series([ts], dtype="datetime64[ns]")
+        expected = Series([ts_naive])
+        tm.assert_series_equal(result, expected)
+
+        with tm.assert_produces_warning(FutureWarning, check_stacklevel=False):
+            result = Series(np.array([ts], dtype=object), dtype="datetime64[ns]")
+        tm.assert_series_equal(result, expected)
+
+        with tm.assert_produces_warning(FutureWarning):
+            result = Series({0: ts}, dtype="datetime64[ns]")
+        tm.assert_series_equal(result, expected)
+
+        with tm.assert_produces_warning(FutureWarning):
+            result = Series(ts, index=[0], dtype="datetime64[ns]")
         tm.assert_series_equal(result, expected)
 
     def test_constructor_datetime64(self):